In Reason before Identity, the published text of his recent Romanes lecture at Oxford, Amartya Sen begins with an anecdote of the kind that has given academics a bad name.  Returning to England on an Indian passport, having given his place of residence as "The Master's Lodge, Trinity College, Cambridge," he was apprehended and asked whether he and the Master were in fact close friends. Though such a misunderstanding were easily rectified, he chose to pause and address the question of whether it were possible that he could be at one and the same time Amartya Sen, Master of Trinity College, and his own close friend. After some debate, he concluded that he is able to ascribe himself both identities, thereby adding to the confusion at immigration. With friends like him, Sen observes, who needs enemies? Nevertheless, his purpose is not to be glib: identity is a vexed issue, and to identify an object only with itself is, given Wittgenstein's caveat, a largely pointless exercise. When we consider the question of identity it is not, we must hope, with a view to coming up with a simple ethnic or political criterion for it, but rather to show that personal identity is complex and that we can ascribe unto ourselves a number of potentially conflicting identities.
The question of identity is, of course, a hot property in the study of the Romantic period. It inheres in the examination of that slippery term 'the self' that some would have us believe has existed only since the counter-enlightenment, and is recurrent in discussions of self-consciousness. To write a poem that is self-consciously poetical is, one may presume, to declare one's identity as a poet. Meanwhile, to write a poetical autobiography is to wrestle with the very act of creating identity. More pertinent to certain recent scholarship have been questions of political or national identity, particularly with regard to the role of the Romantic period in the genesis of modern nationalism and imperialism. As such, identity is very much a matter for scholastic concern. We turn back to the Romantic period as historians in order to understand the origins of European nationalism, or indeed to question those origins, with a view to better understanding the vicissitudes of nationalism in our own century. Alternatively, we look to uncover the sources of a discourse of selfhood that we find alive and well today.
A wholly different, if no less interesting picture is created if our attention, as scholars of the Romantic period, remains closer to home. In short, what, if such a thing exists, does the community of Romantic scholars look like, such that we may both identify ourselves with it, and identify it as a friend? This review-essay is hardly the place to address such a question in full, and in any case it is one that will, I suspect, have been anticipated by anyone who really cares to answer it. A tentative and rather obvious answer may be that Romantic scholarship in its many guises is only an inevitable product of the grip that the multifariousness of the period itself has had upon the very shape of modernity. Still, I begin with this question in all seriousness in part because each of the above studies goes some way to providing an answer to the problems of what is constituted by the legacy of the Romantic period, and of what is meant by the study of the period. The question has, however, a very long history itself, so much so in fact that it may be said to occupy an important place in the debate about the identity and purpose of literary criticism itself, and even of modern identity more generally.
After the preoccupations of New Historicism, recent years have thankfully seen something of a return to formalist concerns in the criticism of the Romantic period. Seamus Perry, reviewing Michael O'Neill's timely Romanticism and the Self-Conscious Poem, heralded such a move in this very journal pointing to the work of Susan Wolfson among others. Those of us lucky enough to have heard Michael O'Neill outline a new project on the theme of inexpressibility at the recent Wordsworth Summer Conference are perhaps mindful of the influence that the work of Paul de Man may have upon any new formalism. His ambiguous presence continues to be felt by critics concerned with close attention to the act of reading. Indeed, de Man has himself been the subject of a recent study by Rodolphe Gasché, The Wild Card of Reading, which builds upon the work of an earlier collection of essays, Reading de Man Reading, thereby suggesting that interest in his work persists.  It certainly does no justice to the work of those critics who are interested in close reading to say that they continue in a de Manian tradition, but such has been the influence of essays like "Shelley Disfigured" that we do well to look again at them.
The path of de Man's career is well known, both in terms of the development of his intellectual interests, and in terms of his fall from grace after extraordinary prominence, following the discovery of his work for the collaborationist press in occupied Belgium during the 1940s. That his political fall coincided with an increase in a historicist agenda and a movement away from the theoretical and comparativist concerns of the Yale School that made his work seem generalist is also readily apparent. Those who have attacked de Man and his followers have suggested that the deconstructive movement was really little more than a new New Criticism, ahistorical in its approach and concerned with method more than scholarship.  The defence undertaken by Gasché assumes that de Man demands respect as a coherent and rigorous thinker whose difficulty is present in his writing itself. It is key to Gasché's argument that de Man combines interests which are philosophical, linguistic and literary, and that find themselves grouped around the question of the identity of rhetoric itself. His argument pays considerable attention to the group of papers that emerged posthumously as Aesthetic Ideology and The Resistance to Theory, and as such his focus is very much upon de Man's role as a philosophical commentator. In Cultural Capital, by contrast, John Guillory addresses de Man's, and indeed deconstruction's, role in the justification of a literary curriculum within the institution of the modern university. He suggests that 'rhetorical reading' is one attempt among many to legitimize literary studies as a 'rigorous' discipline, and is therefore tied inherently to its historical moment. 
In both accounts, the diversity of de Man's interests is implicitly reduced in accordance with the needs of the arguments. Gasché focuses upon the late work preoccupied with the aesthetic philosophy of Kant, Schiller and Hegel, and upon such well-known essays (themselves canonical) as "The Rhetoric of Temporality." Guillory, by contrast, seemingly ignores much of de Man's oeuvre in his fascinating discussion to assess instead the larger project that was deconstruction itself. 'Rhetorical reading' is, as such, reduced in many ways to the kinds of literary criticism that were spawned by the Yale School, such as Cynthia Chase's.  The one ignores the area that the other over emphasizes. Scholars of the Romantic period tend to think of de Man as a critic who was predominantly interested in Romanticism and who altered our reading of central literary texts by focusing upon temporality and indeterminacy. While this is certainly true, the fact is that the corpus of work which deals with writers in the British tradition is extremely small and his impressive theoretical work is largely exhibited in his examination of the European tradition, and most especially of Rousseau. From a methodological point of view, such considerations are all but irrelevant. Nevertheless, such has been the focus upon the contextualisation of the British tradition over the last fifteen years that comparative work on Romanticism now moves tentatively at best. It is perhaps not unreasonable to think that if we wish to refine the concept of Romanticism from the standpoint of the Romantic period in Britain, then de Man may not be our best guide.
Romanticism and Contemporary Criticism seems to announce itself as the most complete statement of de Man's view of the changing conception of identity brought about by the Romantic period, and as such we may well welcome it. The bulk of this collection of essays consists of papers that made up the Gauss Seminar lectures delivered in 1967 at Princeton. They form the groundwork for much of his later work but do not address the question of rhetoric as it emerges in revised form in the later Allegories of Reading. Instead, the focus of the discussion lies in the construction of a Europe-wide and generalized Romanticism that radically changed our understanding of identity. The other essays date from various periods in de Man's career including two late responses to Murray Krieger and Frank Kermode, and an unpublished review of Roland Barthes for the NYRB. All of these are cameo pieces in a collection of academic essays and the Barthes review, especially, seems somewhat misplaced. They do, it is worth noting, provide us with some insight into the extraordinary personal charm that de Man must surely have exuded, and are at once deftly cutting and bashfully self-effacing. The response to Krieger's paper "A Waking Dream: The Symbolic Alternative to Allegory" is instructive as it shows de Man himself in the act of distancing the complexity (as he sees it) of deconstruction from that which he calls, a little patronisingly, the "valuable [...] technical premises of New Criticism." (p. 181) This response, which examines the "Ode to a Nightingale" in the context of Krieger's valorisation of symbol over allegory, is also instructive because it returns us (in 1981) to the territory of "The Rhetoric of Temporality", and demonstrates many of the vicissitudes of de Man's method. He shifts ground rapidly, and before the reader (or auditor) has settled into the argument, she is removed to greater vistas—Krieger to Keats to Descartes, Pascal, Freud, de Man himself. The impression left, as we are set down firm in the conclusion that the poem is the very allegory "of the symbolic, nonaesthetic character of poetic language," (p. 187) is quite simply this: less is more. To be frank, the reader interested in the Romantic period will find little overall that is actually new, and the interest lies instead in the elaboration of de Man's thought from his earlier pre-rhetorical period.
The Gauss Seminar is interesting as a document that dates from the period immediately before "The Rhetoric of Temporality," and that contains the early stages of an ambitious study in which de Man proposed to address the concept of Romanticism in its broadest context. That he did not is perhaps a salutary lesson given the later turn in Romantic studies. There is much of extreme interest here and I shall turn to the text shortly. However, I began with the question of identity not only since this is a series of lectures that addresses the issues of the construction of self-consciousness, but also because of de Man's extraordinary presence in the scholarship of the period in which he was preeminent. It is no exaggeration to say that no major survey of Romantic scholarship before 1990 could fail to acknowledge the presence of de Man and of Yale deconstruction. Indeed, the vast body of criticism that emerged in the wake of deconstruction, in which critics felt comfortable addressing the interrelated concepts of self-consciousness and linguistic indeterminacy across a wide corpus of British and European writing, and in which the discussion was informed by an apparently fixed canon of philosophical and cultural commentators, was as close as Romantic scholarship has come to a single identity within the modern university framework. This would seem far less remarkable were it not for the fact that at the present, a time in which the very teaching of the Romantic period is frequently subsumed within either the extended eighteenth or long nineteenth century, Romantic scholars can seem to share almost no common identity at all.
Behind the rhetoric in de Man's response to Frank Kermode, at the end of this volume, there lurks a note of fear: he rather disparagingly refers to the ranks of non-theoretical critics whom he sees arranging themselves in opposition to his own work. If the battle-lines are being drawn, it is instructive to note that now we would not, as he does, place Kermode himself among them. The state of Romantic scholarship is rather clearly far more fragmented today than even de Man himself imagined it could be. Sites of communal memory are rare indeed in this fragmented world, but one, namely the Lovejoy-Wellek debate, is referred to by de Man in his early essay of 1954, "The Double Aspect of Symbolism." There he refers to the victory scored by Wellek, whom he describes, in a term that may itself raise an eyebrow, as "an orthodox historian": "Until further notice, one is allowed to speak, with reasonable objectivity, of a romantic movement as of a definite and distinctive event in the history of Western literature." (p. 147) It is probably fair to say that Marilyn Butler and Jerome McGann served at notice around the time of De Man's response to Kermode in the form of important publications. If de Man seems dated, I would suggest that it is not (pace John Guillory) as a result of the method that he employs, namely a linguistic and non-historical approach, or not solely because of this. Rather, it is because he firmly adheres to a belief in a single Romantic movement that much modern criticism has sought to present as so multifarious as to be chimerical. David Aram Kaiser puts this point neatly in his study Romanticism, Aesthetics and Nationalism. He contends that while critics like de Man or Abrams saw the British Romantic period as a site of interest because they looked to that which was Romantic (in terms of the 'inward turn' manifest in works like The Prelude), critics like Marilyn Butler, by contrast, were concerned to define that which was precisely British (and different) about it. 
Reading the Gauss Seminar papers is reminiscent of reading Isaiah Berlin on the Romantic period in such essays as "The apotheosis of the Romantic will." I offer the comparison only to suggest that both writers are at their best quite fascinating, and demonstrate the intellectual force of those who are truly at home within the history of ideas. The problem with housing the literary study of the British Romantic period within their domain is that it does not belong there: de Man moves to discuss British writers only quickly to return to the philosophically based European tradition with which he is at home. De Man is really no more likely to influence our reading of some of the less well known figures in the Britain of the 1820s than Berlin, which is really to state that either writer taken on their own terms has much to offer us. It was something of a category error to allow de Man to become the focus of attacks by those who promoted an enlarged canon and a turn from the exclusive 'High Romantic' argument. If a new formalism is to find de Man instructive, it will be as a result of the close attention to the text that he promotes, and this is surely not simply a part of our lost past.
That said, de Man's arguments do frequently sound simply archaic, as when he writes of contemporary criticism as follows:
Contemporary criticism of romanticism has given up attempts at so-called synthesis for the study of particular texts and authors. It has directed its attention toward the texts most likely to contribute to genuine insight and has performed useful revisions in the established hierarchies of romantic authors. It seems now more fruitful, for instance, to approach a poet like Keats or Shelley, or even Coleridge, on the basis of the knowledge one can derive from Wordsworth rather than vice versa—and not for chronological reasons. Not only are [contemporary critics] concerned with the right authors, but they reveal the high level of their understanding by the choice they make of the passages they select for major emphasis.p. 96
It is easy to see that this flippancy may rouse certain readers to anger. Furthermore, de Man, always driven by the pedagogical impulse towards qualitative judgement, prioritizes a canon that represents Romanticism as a historiography of the self. If a 'literary history' of the period cannot be written, then it is because his Romanticism is a diachronic, cross-cultural phenomenon represented to us by a group of texts: The Prelude, Hyperion, Die Phänomenologie des Geistes, Hölderlin's late hymns, Les Fleurs du Mal. For de Man, the investigation of selfhood is the legacy of the period, and we are ourselves living within it, (though one may object that as such it is hard to see where Romanticism ends and the confessional narratives of late twentieth century America begin). Moreover, a near inevitable product of promoting self-consciousness as a criterion for literary excellence is the elevation of Rousseau to a representative figure in the inward turn that de Man marks. In the most substantial of his Gauss lectures, he is concerned to address the relationship between self-reflection and the act of writing, by entering into dialogue with the influential Geneva critic Jean Starobinski. De Man's conclusion points to the linguistic direction in which his own thought is heading:
[Rousseau's] entire effort has been directed toward freeing himself, by reflection, from the burden of his own empirical contingency. In his attempt, he has indeed transcended his actual self into a language, a work that now exists outside himself. [...] But when, as he cannot help himself doing, he starts to reflect upon the work that he has created, he realizes that it only records his failure really to transcend his own selfhood. The work is "encore-moi," the half-resigned, ironic mood of self-reflection that predominates in Rousseau and in the readers who recognise themselves in him.p. 49
The work in question is Rousseau's Pygmalion, but de Man describes the general condition of self-consciousness as it is predicated upon the acts of reading and writing. Furthermore, that Rousseau's self-reflection is also representative of the generalised condition of modernity is to be seen in the responses to his writing that are available to us. If the critic becomes aware of her task in the act of reading Rousseau, (namely to examine patterns of self-conscious reflection), then she is only the latest in a tradition of readers to find her self-awareness enhanced through identification with the author as his is re-created in the text. Literary criticism, as a modern phenomenon, is also the act of searching for the origins of one's imaginative self.
De Man makes this point very nicely in reference to Hazlitt, whose discussion of the Confessions moves quickly into the purely self-conscious genre of recollection. He is then, de Man writes, "reminiscing in a pure Rousseau-like fashion, about the past moment when, in the act of reading Rousseau, he could forget the presence of his own actual identity and lose himself in dreamlike moods of impersonal recollection." (p. 30) So strong was the effect of reading Rousseau, Hazlitt writes, that he has come to realise that "[t]here are, indeed, impressions that neither time nor circumstance can efface." (p. 30) De Man is eager to show that Hazlitt has been so affected by reading Rousseau's self-conscious writing that it becomes impossible to distinguish between the two authors. He does so by demonstrating that the sentence quoted is in fact Rousseau's own. It is a fact that de Man fails to point out that this sentence appears again in Hazlitt, notably when he discusses the reading of Schiller's Die Räuber, though the point hardly militates against him. Nevertheless, in choosing Hazlitt, we have, as it were, a personality tailor-made for de Man as there can hardly be a writer in the period for whom the act of reading is so centrally formative of the self. Moreover, Hazlitt is himself strongly aware of the fact that language can flatter the addressee into taking up false positions, and indeed that in the act of writing we can erroneously suppose ourselves to be asserting an autonomous self that is little more than the product of language use.
If the example of Hazlitt provides an obvious instance of self-formation in the response to the literary work then, as Vincent Newey is keen to show, reading is extendible to something approaching a general principle behind the construction of Romantic autobiography. In an interesting essay on Childe Harold III and IV in Centring the Self, he extends the idea that selves are formed through reading by arguing that Byron reads both natural and cultural objects as though they were texts. It is through the construction of the world as a series of texts, and through the linguistic construction of the self within that world that the self is made as a text for future readers. For Newey, there is a double bind involved in the act of autobiography, for "language itself may delete as well as create 'being'." (p. 179) He has learnt his de Manian lesson well: not only is the self formed through the linguistic response to reading (as in Hazlitt's critical response to Rousseau), but the language of that response takes on a life of its own. Being thus becomes subordinate to the language in which it is uncovered:
Though Childe Harold is vividly referential, impressive in its realizations of history, civilisation, or nature, the 'world' man inhabits, its pressing actuality is not that of any material realm beyond the text but the actuality of the mental process embodied in it, the blend of idealism and vigorous scepticism that is endlessly reconstructive and an endless dissolution, a knowing and not knowing, perception as insight and perception as ignorance, writing and undoing 'what is writ,' finding the signs meaningful or meaningless, being triumphant or defeated.p. 209
In his own way, Newey achieves a rhetoric that is as persuasive as his transatlantic contemporaries, and there seems, in fact, to be little to disagree with here. Nevertheless, one can have the sense that he reads very much within the ideology of the text itself. Surely Byron would wish us to read Childe Harold as an attempt to give coherent significance to subjective experience while acknowledging the artificial and incomplete nature of the task. Perhaps sensing that his reading comes dangerously close to being a little obvious, Newey is at times close to proclaiming a linguistic idealism in which there is nothing beyond the text, and this is really rather unhelpful.
Centring the Self is less a complete study of the theme of self-consciousness in the Romantic period than a collection of essays written over a number of years beginning in the 1970s, each of which finds its theme to be the role of language in the documentation of subjective experience. As such, it is strongly marked by the tenor of the years in which these essays were written. Nevertheless, the central chapters, which deal with Keats, Shelley and Byron, are a fairly coherent entity. The work as a whole hints at on-going themes that relate to the Romantic period by placing these chapters between essays on Cowper, Gray, George Eliot and Thomas Hardy. Of these, the note on Middlemarch is really fairly insubstantial, but the others raise interesting comparisons, most especially a longish piece on the theme of self-reflexivity in Jude, itself a novel that seems to cry out for comparison with Hardy's Romantic forebears. Some readers may indeed find themselves acquainted with a number of these essays from their fore-going publication histories. Of particular interest are the essays on Shelley, which include an interesting reading of "Julian and Maddalo" as a text explicitly concerned with the investigation of personal identity through the theme of doubling, (itself an idea that Newey sees quite rightly to be at the heart of Byron).
Newey appears to see the legacy of Romanticism in the clear dialogue that future writers enter into as they reconsider questions fundamental to the period itself. As such, he moves freely, for example, from a discussion of linguistic indeterminacy to remind us of these lines in Middlemarch: "Signs are small measurable things, but interpretations are illimitable, and in girls of sweet ardent nature, every sign is apt to conjure up wonder, hope, belief, vast as the sky, and coloured by a diffused thimbleful of matter in the shape of knowledge."  Without necessarily adhering to a theory of influence, Romanticism is seen as such to be diachronic phenomenon, identified with 'ways of thinking.' As I have remarked, much of de Man's work can be read as a dialogue conducted with other significant critics on precisely this subject, (though he treats conversation somewhat as if it were a game to be won at all costs). It is this dialogic aspect of literary study that Orrin Wang uses as the basis for his interesting study Fantastic Modernity. I think that it is fair to say that this is potentially a highly significant work, though in tackling the relationship between theories of Romanticism and political practice, specifically as they are construed in postmodernity, it is unambiguously ambitious.
Wang takes as his starting point the fact that Romanticism, as a cultural phenomenon, has always been mediated, or more precisely has always been a creation of its appropriators. It has, furthermore, been constructed so as to be dismissed by the major mediating force of postmodernity, namely high modernism. Wang points out that while the study of Romanticism was returned to the literary curriculum by a group of critics who were politically quietist, it has been appropriated in late modernity by disparate schools of criticism for whom the concerns of the Romantic period are more current than our modernist forebears. As such, study of the Romantic period has been dominated in recent years by those interested in the critique of ideology, by feminists and new historicists concerned with overturning conventional histories of modernity, and by those fascinated by linguistic indeterminacy. Wang is interested not only in the identity of Romantic studies such that it may be able to house so apparently disparate a group of bed-fellows, but also with the contradictions inherent in post-Romantic modernity that enable us to turn Romanticism to such very different ends. In chapters on McGann and Heine, de Man and Shelley, Bloom and Emerson, and in general discussions of feminism and of the debate about Romanticism's heritage in Marxism and Fascism, Wang pairs a series of representative critics alongside significant writers and works in order to investigate relevant critical methodologies. His aim is not only to show how texts have been turned to specific ends, but also that the texts themselves inform upon the very general contradictions that he points to.
Essentially, Wang is asking whether post-enlightenment modernity can be reduced to a single historical entity, or whether, in devising terms like Romanticism and post-modernity, we are simply coming up with different ways of naming parts of its heterogeneity. Throughout he is employs a dialectical method for understanding this relationship, exhibited in his practice of reading Romantic texts alongside their commentators. Nevertheless, this project insists upon a critical language that is often confusing and complicated. The following passage amply demonstrates the complexity of his task, and the need for an equivalently taxing effort of the part of the reader:
This book does not [...] focus upon Romanticism as simply an originary influence upon critical theory, in which a return to the Biographia Literaria might unproblematically explain the tenets of either New Criticism or deconstruction. Neither does this book solely view Romantic culture as a hegemonic formation that founds all the institutional and social conditions of intellectual labor that follow it. In that sense this book acknowledges the mediated nature of Romanticism as an-always-already theorized concept, yet it also insists upon the mediated state of theoretical "metacommentary" that delivers Romanticism to us. To begin to understand the social, cultural, and epistemological consequences of such metacommentary - to denaturalise and expose the commentary's own contingencies - a dialectical reading returns to the late-eighteenth and early-nineteenth century texts that retroactively form, though neither simply nor absolutely, Romantic discourse. Romanticism thus functions in this book as something besides either a completely mysified retroactive construction or a completely transparent, originary form of knowledge. The former is too determined, the latter too deterministic. Hence my stress on the mutually transforming dialectic that binds Romanticism and theory together. Romanticism is the transcoding of what it itself transcodes: at this moment, most visibly and forcefully, the discourse of postmodern theory.p. 10
As one wrestles with this study and becomes attuned to what Wang is getting at, it becomes clear that his method makes good sense, and is a actually remarkably clever way of unpacking the vicissitudes of recent scholarship. Far too much attention has been paid either to attempting 'new' ways of reading the works known to us, or to attacking the scholastic methods employed by a host of whipping-boys.
The use of the word 'transcode' is perhaps the key, not least because it is initially unclear what Wang is getting at. Neither romanticism nor contemporary scholarship exist as independent entities, nor is one simply the construct of the other. Both are somehow inherently indeterminate, and the clear fact is that the world we inhabit, as social and political beings, is apart from either one. For Wang, the act of the critic remains relevant only so long as theory is tied to practice, or as long as the reading of texts relates to our social and political being. If he is largely concerned with the interpretation of politicised texts , then this is in order to show how such texts, as ambiguous political statements, may be turned to specific ends in the hands of the critic. 'Transcoding' is a fine description of this endeavour. To look, for example at the use to which de Man puts Shelley's "The Triumph of Life," is to examine the way in which the linguistic indeterminacy of the text is read as a determined condition of postmodern existence. Wang, by contrast, is eager to show that the quality of the text that de Man ignores, namely its complex political essence, informs upon the circumstances of de Man's own reading. To place the reading alongside the text is to show that 'Romantic' and 'Postmodern' are both the unsettled quantities, which can only be approached by reading both as codes alongside one another. Romanticism as a scholarly construct is a term which describes our attempt to understand how a large group of different men and women attempted in turn to give significance to the private, social and political worlds that they inhabited, but the attempt to describe their identity accurately is rather clearly only part of an attempt to describe our own. If, at the end of the century, this seems to be a truism, then it is worth noting that it was not always so.
Amartya Sen, Reason before Identity (Oxford: Oxford university Press, 1999) p. 1.
Roldophe Gasché, The Wild Card of Reading: On Paul de Man (Boston: Harvard University Press, 1998), and Lindsay Walters and Wlad Godzich, eds., Reading de Man Reading (Minneapolis: University of Minnesota Press, 1989).
See Chaviva Hosùek and Patricia Parker, eds., Lyric Poetry: Beyond New Criticism (Cornell: Cornell University Press, 1985) p. 346.
John Guillory, Cultural Capital: The Problem of Literary Canon Formation (Chicago: Chicago University Press, 1993) pp. 260ff.
See Cynthia Chase, Decomposing figures: Rhetorical Readings in the Romantic Tradition (Johns Hopkins University Press, 1986).
For Abrams, English Romanticism is romantic because it shares the worldview of German idealism; for Butler, English Romanticism is English precisely because it does not [David Aram Kaiser, Romanticism, Aesthetics and Nationalism (Cambridge: Cambridge University Press, 1999) p. 13].
George Eliot, Middlemarch, ed. David Carroll (Oxford: Oxford University Press, 1997) p. 23.